← All publications
Automated LinkedIn Publishing with Python Cron 6/6
15 posts. 2 languages. Scheduled over 13 days. Zero manual intervention after validation.
Here's the automated publishing system.
🔑 Step 1 — Create a LinkedIn app
1. linkedin.com/developers → Create App
2. Link your company page (required)
3. Request scopes: openid, profile, w_member_social
4. Redirect URI: http://localhost:8099/callback
5. Repeat for each LinkedIn profile
🐍 Step 2 — OAuth authentication script
A Python script starts a local HTTP server on port 8099. You open the authorization URL in your browser. LinkedIn returns a code. The script exchanges it for a token valid for 60 days.
I have 2 profiles: FR (hicommerceweb@gmail.com) and EN (stephane.jambu@gmail.com). Each has its own .env file and token.json.
📤 Step 3 — Publishing script
The post.py script reads a text file and publishes to LinkedIn via the REST API:
• POST https://api.linkedin.com/rest/posts
• Required header: LinkedIn-Version: 202502
• Options: --profile fr|en, --file post.txt, --image photo.jpg
Usage: python post.py --profile fr --file post.txt
Limit: 3,000 characters per post.
⏰ Step 4 — Scheduling with cron
Each post has an FR and EN file (translated by Claude Code — not a literal translation: currencies adapted, tone adjusted).
A cron job triggers publishing at scheduled times:
• FR at 9:30 AM France time
• EN 6 hours later (3:30 PM France = 9 AM US Eastern)
A bash wrapper handles logs and errors.
💰 Total system cost
• LinkedIn API: free
• Server: none (cron runs on my PC)
• Writing: Claude Code (existing subscription)
• Additional cost: $0
🔄 The complete cycle
Zoom meeting → auto transcript → AI summary in Obsidian → insight extraction → post writing → human validation ("OK") → auto scheduling → FR+EN publishing
From meeting to LinkedIn post, the only moment I step in: I read the post and type "OK".
Everything else is automated.
Are you still publishing manually on LinkedIn? How much time do you spend on it each week?
Here's the automated publishing system.
🔑 Step 1 — Create a LinkedIn app
1. linkedin.com/developers → Create App
2. Link your company page (required)
3. Request scopes: openid, profile, w_member_social
4. Redirect URI: http://localhost:8099/callback
5. Repeat for each LinkedIn profile
🐍 Step 2 — OAuth authentication script
A Python script starts a local HTTP server on port 8099. You open the authorization URL in your browser. LinkedIn returns a code. The script exchanges it for a token valid for 60 days.
I have 2 profiles: FR (hicommerceweb@gmail.com) and EN (stephane.jambu@gmail.com). Each has its own .env file and token.json.
📤 Step 3 — Publishing script
The post.py script reads a text file and publishes to LinkedIn via the REST API:
• POST https://api.linkedin.com/rest/posts
• Required header: LinkedIn-Version: 202502
• Options: --profile fr|en, --file post.txt, --image photo.jpg
Usage: python post.py --profile fr --file post.txt
Limit: 3,000 characters per post.
⏰ Step 4 — Scheduling with cron
Each post has an FR and EN file (translated by Claude Code — not a literal translation: currencies adapted, tone adjusted).
A cron job triggers publishing at scheduled times:
• FR at 9:30 AM France time
• EN 6 hours later (3:30 PM France = 9 AM US Eastern)
A bash wrapper handles logs and errors.
💰 Total system cost
• LinkedIn API: free
• Server: none (cron runs on my PC)
• Writing: Claude Code (existing subscription)
• Additional cost: $0
🔄 The complete cycle
Zoom meeting → auto transcript → AI summary in Obsidian → insight extraction → post writing → human validation ("OK") → auto scheduling → FR+EN publishing
From meeting to LinkedIn post, the only moment I step in: I read the post and type "OK".
Everything else is automated.
Are you still publishing manually on LinkedIn? How much time do you spend on it each week?